Submitted by admin on Thu, 08/18/2022 - 23:47

Lack of video surveillance and AI recognition

Sequence Number 2 Industry Seaports Banner Lack of video surveillance and AI recognition How 5G enabled

Intelligent analysis of operators’ facial expressions and status with alarms for fatigue and sleepiness in operation management along with license plate recognition and facial recognition for security.

Data Flows
Title Devices Icon Devices Description
  • Sensors (cameras) located at multiple locations across the site: fixed/drones/robotics/cars/staff/etc.
  • Timeseries data to support camera images
Title Connectivity Icon Connectivity Description
  • Time series data transport
  • Camera images
  • Asset data (maintenance records) access
Title Edge Compute Icon Edge Compute Description
  • Several activities are real-time activities
  • Camera – MV interpretation
Title Cloud Compute & Storage Icon Cloud Compute & Storage Description
  • All data collected from assets (both historical and real-time)
  • Enterprise-owned storage
Title Applications & Services Icon Applications & Services Description
  • Non-time critical activities
  • MV and some ML focus; E2E automated
  • Multiple MV models linked to apps
Title Inform Decision Makers Icon Inform Decision Makers Description
  • Errors and safety violations reported immediately to operations centre of the site 
Title Support Decision Making Icon Support Decision Making Description
  • End of process
Application Logic
Description
  • List the business cases to be tracked: Intelligent analysis of operators' facial expressions and status with alarms for fatigue and sleepiness operation management, license plate recognition, and facial recognition.
  • Collect camera images (fixed/drones/robotics /etc.) from any potential source with minimum image quality of 720p.
  • Collect as much data as possible about selected assets and surroundings as ML models will be strengthened with more data.
  • Edge + 5G to be used for all time critical events.
Description
  • All data collected from edge sensors will be stored long-term in Enterprise storage.
  • Focus on MV to identify data needed for the different scenarios.
  • Development of the ML model is done through an iterative process and a quality ML model (fully data-driven) will require multiple steps to detect anomalies/potential failures.
  • Models will be stored and maintained by AI applications.
  • SME involvement working with data scientists is required to develop the model.
Description
  • The process from data collection to execution of models is fully automated.
  • MV models developed to support all application cases.
  • Various options for visualisation of entire operations (e.g., XR options on OpenXR platform).
Expected benefits Key value created